Factors that Increase the Likelihood of Ovarian Endometriosis Relapsing after Surgical Excision: A Systematic Review and Meta-Analysis

Background : Determine potential triggers for ovarian endometriosis recurrence to provide individualized long-term management and follow-up for improved patient outcomes. Methods : Relevant data were acquired through systematic retrieval from PubMed, Embase, Web of Science, and Cochrane Library before October 2022. We determined the odd ratio or the mean difference with their corresponding 95% confidence interval (CI) to explore the relationship between relevant risk factors and postoperative endometriosis relapse. Results : This meta-analysis ultimately covered 6388 patients from 18 trials, and the findings demonstrated that postoperative endometriosis recurrence was associated with the age at surgery [mean difference (MD) (95% CI) = –0.69 (–1.33~–0.05), p = 0.04], family history [odds ratio (OR) (95% CI) = 2.18 (1.10~4.29), p = 0.02], preoperative carbohydrate antigen (CA-125) [MD (95% CI) = 24.08 (–7.55~40.61), p = 0.004], laterality of endometriosis (EMs) [OR (95% CI) = 1.19 (1.00~1.40), p = 0.04], presence of adenomyosis [OR (95% CI) = 1.53 (1.11~2.11), p = 0.009], presence of myoma [OR (95% CI) = 1.44 (1.07~1.94), p = 0.02], previous endometriosis-related surgery [OR (95% CI) = 1.90 (1.45~2.51), p < 0.00001], and r-American Fertility Society (r-AFS) stage [OR (95% CI) = 0.30 (0.19~0.46), p < 0.00001]/[OR (95% CI) = 0.57 (0.48~0.66), p < 0.00001]. In addition, postoperative pregnancy [OR (95% CI) = 0.40 (0.19~0.82), p = 0.01] and postoperative medication [OR (95% CI) = 1.64 (1.02~2.62), p = 0.04] were indicated protective factors for the prevention of postoperative ovarian endometriosis relapse. Conclusions : Risk factors for postoperative endometriosis recurrence included the age at surgery, family history, CA-125, laterality of EMs, presence of adenomyosis, presence of myoma, previous endometriosis-related surgery, and r-AFS stage. In addition, protective factors for preventing postoperative recurrence included postoperative pregnancy and postoperative medication. However, the age of menarche, infertility, the extent of surgery, cyst size, body mass index, and dysmenorrhea were unrelated to postoperative recurrence. The sample size could be increased for further investigations.


Introduction
Endometriosis is a common disease affecting women of childbearing age, where the endometrium tissue (both interstitial and glandular) grows out of the uterine cavity. The periodic bleeding from the ectopic endometrium and fibrosis of the surrounding tissue often form ectopic nodules and other lesions. Pathologically, endometriosis presents benign morphological changes but invasive, metastatic, and recurrent biological behaviors of malignant tumors, earning it the moniker of benign cancer. The major symptoms include dysmenorrhea, chronic pelvic pain, irregular menstruation, and infertility [1]. Endometriosis can occur in all pelvic tissues and organs, particularly the ovaries [2].
Surgery is the standard treatment for ovarian endometriosis, but the postoperative recurrence rate is high. Recurrent endometriosis is the regeneration or reappearance of lesions after standard surgery and drug treatment, with clinical symptoms returning to or even exceeding the level before treatment after symptom relief. According to previous studies, the five-year recurrence rate after conser-vative surgery is over 40% [3]. The high postoperative recurrence rate severely affects patients' physical and mental health and quality of life [4,5] and increases their social and economic burden, making it a challenging issue clinically. Preventing or delaying postoperative recurrence of ovarian endometriosis is a difficult problem to tackle. However, the pathogenesis of endometriosis has not been fully elucidated, and no single theory can explain its occurrence [6]. Therefore, it is necessary to determine the factors associated with its clinical recurrence rate.
This study collected data on relapse-related factors from long-term follow-ups of childbearing-age patients with ovarian endometriosis. Assessing the association between these factors and ovarian endometriosis recurrence to identify the risk factors may help to provide individualized long-term management and follow-up based on clinical conditions and reduce postoperative recurrence.

Inclusion and Exclusion Criteria 2.2.1 Inclusion Criteria
(1) Studies of factors related to the postoperative recurrence in patients with ovarian endometriosis; (2) Cohort studies or case-control studies; (3) With subjects concerning patients with ovarian endometriosis confirmed by postoperative pathological exam-ination; (4) With a postoperative follow-up time of at least six months; (5) The postoperative recurrence was confirmed by ultrasound and reoperation.

Data Extraction and Quality Assessment
With the search terms and inclusion and exclusion criteria, relevant studies were retrieved from the above databases. The relevant literature was initially screened by reading the titles and abstracts, and those that failed to meet the requirements were excluded. Then, the full texts were read, and those that failed to meet the requirements were further excluded. Disagreements on literature inclusion were resolved through discussion or judgment by a third researcher. Studies with similar contents were carefully checked in terms of the title, author, and affiliation to determine whether they were duplications. Only one of the confirmed duplications was included.

Statistical Analysis
Review Manager 5.4 (the International Cochrane Collaboration, Nordic Cochrane Centre, Oxford, United Kingdom) was employed for the analysis. Efficacy analysis statistics were expressed by odds ratio (OR) for dichotomous variables and by mean difference (MD) and 95% confidence interval (CI) for continuous variables to evaluate the impact of related factors on recurrence. The test level was p < 0.05, and the I 2 statistic was used to assess heterogeneity. The fixed effects model was adopted since homogeneity was indicated by p > 0.1 and I 2 ≤50%. If p ≤ 0.1 and I 2 >50%, the random effects model was used considering the large heterogeneity among studies. Sensitivity analysis was performed to assess the stability and reliability of the pooled results. Publication bias was evaluated using funnel plots.

Literature Retrieval
The literature retrieval process yielded 6768 potentially eligible studies. In addition, 3 studies were identified from the reference lists of the included studies. A total of 4615 studies were screened out by two reviewers independently after excluding duplicate studies. After reading the titles and abstracts, 4542 studies not associated with recurrent risk factors for ovarian endometriosis were further excluded. The remaining 73 studies were then fully reviewed. Among them, 5 were excluded due to insufficient data; 3 studies were excluded due to inconclusive postoperative pathologic findings; 25 studies were disqualified due to the lack of a control group or a non-recurrent group as the control group; 19 studies were excluded as cyst recurrence was not selected as an outcome; and 14 studies were disregarded due to insufficient postoperative follow-up. Finally, this meta-analysis included 18 studies satisfying the inclusion criteria. The flow chart of the literature selection process is shown in Fig. 1.
in Table 4. The analyzed potential risk factors included age of menarche, age (years), infertility, postoperative pregnancy, BMI (kg/m 2 ), preoperative CA-125 (U/mL), family history, dysmenorrhea, r-AFS stage, presence of adenomyosis, presence of myoma, cyst size, laterality of EMs, cystectomy of the affected site, postoperative medication, and previous endometriosis-related surgery. With 18 studies having NOS scores between 6 and 8, the included studies were of moderate to high quality (Tables 1,2).
Age of menarche: A total of 3 studies [20,23,24] reported the connection between the age of menarche and postoperative recurrence (I 2 = 0%, p = 0.72). With little heterogeneity among the 3 studies, the fixed-effects model was used, and the results revealed no significant difference in the age of menarche between the recurrence group and the non-recurrence group (MD = -0.18, 95% CI = -0.44~0.08, Z = 1.35, p = 0.18; Fig. 2B).
Endometriosis operation history: There were 4 studies [8,10,14,16] investigating the effect of previous relevant surgery history on postoperative endometriosis recurrence. Due to the little heterogeneity among the 4 studies (I 2 = 45%, p = 0.14), the fixed-effects model was applied. The results revealed that previous relevant surgery history significantly affected the recurrence, and the recurrence rate was high in patients with previous related surgeries (OR = 1.90, 95% CI = 1.45~2.51, Z = 4.58, p < 0.00001; Fig. 2F).
Campo et al. [21] 2014 The Department of Obstetrics and Gynecology of the Catholic University of the Sacred Heart Retrospective Study (1) At least 1 ovarian endometrioma >2 cm in diameter.
He et al. [  (2) Conducted ultrasonography to determine endometrioma recurrence at least 6 months after surgery; Zhang et al. [13] 2019 The (2) Ultrasonography showed that there was growth and infiltration outside the endometrium, and repeated bleeding formed nodules and masses, which caused pain;  (2) Underwent total laparoscopic cystectomy for ovarian endometriosis with or without excision of deep endometriosis (DE).
r-AFS stage: There were 6 studies [7,10,12,13,19,22] that reported on the difference between stages I and II versus stages III and IV on endometriosis relapse, which had little heterogeneity (I 2 = 43%, p = 0.12). The fixed-effect model was used to reveal the effect of the r-AFS stage on postoperative recurrence. The recurrence rate of stage I and II patients was 0.30 compared to stage III and IV patients (OR = 0.30, 95% CI = 0.19~0.46, Z = 5.46, p < 0.00001; Fig. 2O).

Sensitivity Analysis
Sensitivity analysis was performed on postoperative pregnancy, infertility, combined myoma, combined adenomyosis, endometriosis operation history, ovarian cyst size, BMI, and r-AFS stage. After removing the data one by one, no significant change was found in the analysis results, indicating the reliability of the summary results. Sensitivity analysis was also performed on dysmenorrhea. Following the exclusion of the study by Wacharachawana et al. [16], the heterogeneity was statistically changed (I 2 = 44%, p = 0.13).
The results revealed that dysmenorrhea had an effect on endometriosis recurrence, and the recurrence rate of patients with dysmenorrhea was high (OR = 2.44, 95% CI = 1.26-4.73, Z = 2.65, p = 0.008).

Publication Bias
r-AFS stage was used for the funnel plot analysis to assess the publication bias. As shown in Fig. 3, all the included studies are in inverted funnel plots with good symmetry. Therefore, the publication bias of the included studies has relatively little impact on the results.

Discussion
As a benign disease, the exact pathogenesis of endometriosis has not been fully elucidated. The numerous theories include endometrial implantation theory [25], lymphatic and venous dissemination theory [26], coelomic metaplasia theory [27], and immunity and inheritance theory [28]. However, it is difficult to explain the exact mechanisms of endometriosis with one theory, and many hypotheses are linked to the pathogenesis and progression of endometriosis. Currently, the most widely accepted theory of endometriosis is endometriotic implantation caused by retrograde menstruation, in which endometrial glandular epithelial and mesenchymal cells are ectopically implanted in the ovary by retrograde menstruation, where they continue to grow and spread to form ectopic lesions [6]. Endometriosis does not, however, occur in all women with retrograde menstruation. The hypothesis of abnormal eutopic endometrium partly explains why most women of childbearing age have retrograde menstruation and only a few develop endometriosis, which constitutes an important supplement and extension to the theory of ectopic implantation [29,30]. This hypothesis may provide a satisfactory explanation for the same mutations carried by the true and ectopic endometrial tissues in patients. However, in the study by Koppolu et al. [31], no variants were found common to both tissues from the same patient, thus not supporting this hypothesis.
Recurrence is one of the many difficulties in the clinical management of endometriosis. Identifying the factors related to postoperative endometriosis recurrence can provide insights into the individualized treatment of endometriosis to help clinicians determine the risk of recurrence and formulate treatment plans from the perspective of personalized medicine.

Risk Factors for Endometriosis Recurrence
Age at surgery: Studies have revealed that the endometriosis recurrence rate decreases with age, while the age at the time of surgery is a risk factor for recurrence. Higher estrogen levels in young women induce persistent growth of residual lesions. The study by Yang et al. [32] also demonstrated that younger age at surgery was a significant risk for postoperative recurrence of ovarian endometriosis. Therefore, more attention should be attached to young female patients, and the number of follow-up visits should be increased.
Family history: This study suggests that family history has some impacts on postoperative endometriosis recurrence and confirms the existence of genetic trait changes in the pathogenesis of endometriosis. The genetic differences lead to molecular biological differences between the eutopic endometrium of patients and normal subjects.
Preoperative CA-125: Carbohydrate antigen CA-125 is a glycoprotein derived from coelomic epithelial cells and expressed in normal tissues, which could offer some insights into the diagnosis of endometriosis [33]. The above studies concluded that postoperative CA-125 was a risk factor for postoperative endometriosis recurrence. However, CA-125 is a sensitive but poorly specific indicator, and many other diseases also cause marked increases in CA-125. Therefore, CA-125 can be used as a follow-up index but not a diagnostic index of postoperative recurrence. In addition, CA-125 levels can vary with the menstrual cycle. An increased sample size is needed to improve the accuracy of this study.

Laterality of EMs:
This study revealed that the recurrence rate of patients with bilaterality of cysts was significantly higher than that of patients with unilaterality. It is speculated that bilateral cysts may increase the possibility of residual lesions. This conclusion is consistent with the current clinical consensus but contrary to the conclusion of Han et al. [34] and Koga et al. [35], suggesting no significant association between bilateral cysts and postoperative recurrence. Therefore, it is necessary to expand the sample for further study.

Presence of adenomyosis:
The results revealed that adenomyosis was also a risk factor for postoperative endometriosis recurrence as the pathogenesis of endometriosis and adenomyosis overlap [36,37]. Therefore, the correct identification of endometriosis combined with adenomyosis and the formulation of personalized long-term postoperative management plans play a crucial part in recurrence prevention.
Presence of myoma: Fibroids in the uterus were regarded as estrogen-dependent tumors. An animal model of fibroid xenotransplantation revealed that steroids, including estradiol and progesterone, were necessary for tumor growth [37]. Therefore, the high estrogen in patients with uterine fibroids affects the endometriosis recurrence rate. Clinically, long-term management plans and regular follow-ups should be made for postoperative patients with uterine fibroids to prevent recurrence.

Previous endometriosis-related surgery:
The results suggested that patients with previous endometriosis surgeries had a higher postoperative recurrence rate, possibly because of the more aggressive endometriosis type in patients with previous related surgeries. The mainstream view is that whether the lesion is completely removed during the operation determines the postoperative recurrence. Therefore, surgeons should remove the lesion as much as possible during the operation rather than relying on postoperative adjuvant drugs. Radical surgery is recommended for patients at high risk of recurrence and without fertility requirements.
r-AFS stage: r-AFS stage is determined based on endometriosis lesion size, location, degree of adhesion, etc. [1,38]. This study concluded that the postoperative recurrence rate was higher with higher surgical scores and more severe stages. In short, extensive lesion adhesions during surgery tend to cause residual lesions and increase the risk of recurrence. Therefore, through clinical analysis, the r-AFS stage is of certain reference significance to evaluate the prognosis and disease severity. When selecting treatment plans, the patient's r-AFS stage should be considered.

Protective Factors for Endometriosis Recurrence
Postoperative pregnancy: The combined data analysis confirmed the benefit of postoperative pregnancy in preventing endometriosis recurrence, possibly because pregnancy and postpartum lactation can lead to ovarian nonovulation and endometrium decidual-like changes, causing the necrosis of endometriosis lesions and reducing the postoperative recurrence rate [39]. Therefore, Vercellini et al. [40] also found that after endometriosis surgery, patients with fertility requirements were preferably impregnated sooner.

Postoperative medication:
The integrated data revealed that patients treated with medication had a lower risk of endometriosis recurrence. Gonadotropin-releasing hormone agonist (GnRHa) can effectively inhibit the secretion of endogenous gonadotrophins and lead to a hypoestrogenic state in vivo, causing deciduosis of endometrium and necrosis of endometriosis lesions [41]. The study by Adachi et al. [42] showed that the endometriosis recurrence rate was significantly reduced after 24 months of dienogest (DNG) administration. In recent years, as an important means of long-term maintenance management after surgery, drug therapy has become a common clinical treatment. In the meantime, Chiu et al. [43] also showed that postoperative adjuvant plus maintenance therapy might be the most effective intervention to prevent endometriosis recurrence. However, the variety of drugs and the complex action mechanisms require more accurate and extensive studies to clarify the efficacy and safety of drugs [44].

Unrelated Factors for Endometriosis Recurrence
Age of menarche: Ponomarenko et al. [45] concluded that some single nucleotide polymorphisms (SNPs) for the age of menarche were associated with endometriosis. Meanwhile, Vercellini et al. [40] found in their study on endometriosis that the early age of menarche in modern women was closely related to the high incidence of endometriosis. However, this meta-analysis suggested that the age of menarche had no significant effect on postoperative endometriosis recurrence. The increased incidence of contemporary endometriosis may be attributed to the improved detection rate. The understanding of the disease is gradually deepened, and the disease becomes treatable, thus increasing the detection rate of endometriosis.

Infertility:
The mechanisms by which endometriosis induces infertility are multifactorial, involving mechanical, molecular, genetic, and environmental causes. New research has found that changes in gene expression and gene defects are also responsible [46]. However, infertility could be caused by a variety of factors, the most common one being female pelvic inflammatory diseases, which cause the fallopian tubes to adhere to prevent the ovum and sperm from bonding and have no significant association with en-dometriosis incidence. Therefore, the fallopian factor accounting for a large proportion of infertility may affect the accuracy of the study results.
Cystectomy of the affected site: It is generally accepted in clinical practice that lesion resection alone leads to an increased likelihood of endometriosis recurrence. However, the combined evidence here suggested that postoperative recurrence rates were not significantly increased in patients with lesion resection alone.
Cyst size: Jiang et al. [47] revealed that greater cyst diameters led to higher possibilities of endometriosis recurrence. On the one hand, the intraoperative residual lesions of large cysts are more likely to conceal the potential factors for recurrence. On the other hand, large cysts often represent greater ectopic cell vitality, thus increasing the recurrence rate. Nevertheless, the findings revealed no association between cyst size and recurrence. We concluded from the overall review that the possible cause was the lack of an objective standard for the intraoperative cyst size evaluation, which was mainly based on the subjective judgment of the surgeon, thus bringing errors to cyst size calculation. Therefore, improving the objective evaluation criteria for cyst size is suggested to increase the accuracy of results.
BMI: As the endometrial tissue grows outside the uterine cavity, the progesterone and estrogen signals are interfered with, leading to progesterone resistance and estrogen dominance in patients with endometriosis [27,48]. Hormone disorders may lead to obesity in the patients and thus increased BMI. It can be inferred that BMI may be related to the recurrence of endometriosis. The study of the relationship between ovarian endometriosis and obesity here revealed no significant association between BMI and endometriosis recurrence.
Dysmenorrhea: Previous studies have shown that the direct and indirect effects of local bleeding in endometriosis tissue, the actions of celiac inflammatory cytokines, and stimulation or direct infiltration of pelvic floor nerve may be the causes of dysmenorrhea [49]. Meanwhile, patients with dysmenorrhea generally have more extensive lesions, which are difficult to completely remove by surgery, thus explaining the increased recurrence rates in patients with dysmenorrhea. However, this analysis suggested that dysmenorrhea was not a risk factor for recurrent endometriosis, contrary to previous findings by Chon et al. [22]. After removing the study by Wacharachawana et al. [16] with heterogeneous source, we found that patients with dysmenorrhea had a higher recurrence rate after surgery (OR = 2.44, 95% CI = 1.26~4.73, Z = 2.65, p = 0.008), which was consistent with the research findings of Chon et al. [22]. After carefully studying the research by Wacharachawana et al. [16], no cause for heterogeneity was found. Therefore, an increased sample is crucial for further exploration to enhance the reliability of the results.

Conclusions
In summary, the pooled evidence suggested that the following risk factors were significantly associated with endometriosis recurrence, namely, age of surgery, family history, preoperative CA-125, laterality of EMs, presence of adenomyosis, presence of myoma, previous endometriosisrelated surgery, r-AFS stage, postoperative pregnancy, postoperative medication. The age of menarche, infertility, the extent of surgery, cyst size, BMI, and dysmenorrhea were not associated with endometriosis recurrence. The risk factors for relapse could be identified, and individualized long-term management and follow-up could be tailored to patient conditions. The sample size can be increased for large-scale studies to confirm the inevitable limitations of this meta-analysis.

Availability of Data and Materials
All data generated or analyzed during this study are included in this published article.